diff --git a/wwwroot/knowledge_bases_list/detail.ui b/wwwroot/knowledge_bases_list/detail.ui index 64f2904..d0109cb 100644 --- a/wwwroot/knowledge_bases_list/detail.ui +++ b/wwwroot/knowledge_bases_list/detail.ui @@ -6,10 +6,9 @@ "widgettype": "HBox", "options": {"width": "100%", "bgcolor": "#f8f9fa", "padding": "10px 16px", "alignItems": "center", "borderBottom": "1px solid #e0e0e0"}, "subwidgets": [ - {"widgettype": "Text", "options": {"text": "← 知识库列表", "css": "clickable", "cfontsize": 14, "color": "#4a90d9"}, "binds": [ - {"wid": "self", "event": "click", "actiontype": "urlwidget", "target": "app.rag_main_content", "mode": "replace", - "options": {"url": "{{entire_url('/rag/knowledge_bases_list/index.ui')}}"}} - ]}, + {"widgettype": "Text", "options": {"text": "← 知识库列表", "css": "clickable", "cfontsize": 14, "color": "#4a90d9"}, + "binds": [{"wid": "self", "event": "click", "actiontype": "urlwidget", "target": "app.rag_main_content", "mode": "replace", + "options": {"url": "{{entire_url('/rag/knowledge_bases_list/index.ui')}}"}}]}, {"widgettype": "Text", "options": {"text": " / ", "cfontsize": 14, "color": "#ccc"}}, {"widgettype": "Text", "options": {"text": "📚 {{params_kw.kb_name or '知识库'}}", "cfontsize": 16, "fontWeight": "bold", "color": "#333"}}, {"widgettype": "Text", "options": {"text": "", "css": "filler"}} @@ -21,63 +20,58 @@ "subwidgets": [ { "widgettype": "VBox", - "options": {"cwidth": 16, "bgcolor": "#fafafa", "borderRight": "1px solid #e0e0e0", "spacing": 0}, + "options": {"cwidth": 16, "bgcolor": "#fafafa", "borderRight": "1px solid #e0e0e0", "spacing": 0, "overflow": "hidden", "flexShrink": "0"}, "subwidgets": [ - { - "widgettype": "HBox", - "options": {"padding": "10px 12px", "bgcolor": "#f0f0f0", "borderBottom": "1px solid #e0e0e0"}, - "subwidgets": [ - {"widgettype": "Text", "options": {"text": "📁 目录结构", "cfontsize": 13, "fontWeight": "bold", "color": "#666"}} - ] - }, - { - "widgettype": "Tree", - "options": { - "id": "dir_tree", - "parentField": "parentid", - "idField": "id", - "textField": "label", - "dataurl": "{{entire_url('./get_tree_data.dspy')}}?kb_id={{params_kw.kb_id}}", - "newdata_params": {"kb_id": "{{params_kw.kb_id}}"}, - "editable": { - "fields": [ - {"name": "name", "title": "名称", "type": "str", "length": 255, "uitype": "str", "label": "名称"} - ], - "add_url": "{{entire_url('./new_tree_item.dspy')}}", - "update_url": "{{entire_url('./update_tree_item.dspy')}}", - "delete_url": "{{entire_url('./delete_tree_item.dspy')}}" - } - } - } + {"widgettype": "HBox", "options": {"padding": "10px 12px", "bgcolor": "#f0f0f0", "borderBottom": "1px solid #e0e0e0"}, + "subwidgets": [{"widgettype": "Text", "options": {"text": "📁 目录结构", "cfontsize": 13, "fontWeight": "bold", "color": "#666"}}]}, + {"widgettype": "Tree", "id": "dir_tree", + "options": {"select_only": true, "title": "目录", "parentField": "parentid", "idField": "id", "textField": "label", + "dataurl": "{{entire_url('./get_tree_data.dspy')}}?kb_id={{params_kw.kb_id}}", + "newdata_params": {"kb_id": "{{params_kw.kb_id}}"}, + "editable": {"fields": [{"name": "name", "title": "名称", "type": "str", "length": 255, "uitype": "str", "label": "名称"}], + "add_url": "{{entire_url('./new_tree_item.dspy')}}", + "update_url": "{{entire_url('./update_tree_item.dspy')}}", + "delete_url": "{{entire_url('./delete_tree_item.dspy')}}"}}} ] }, { "widgettype": "VBox", "id": "detail_content", - "options": {"css": "filler", "padding": "16px", "spacing": "0"}, + "options": {"css": "filler", "padding": "16px", "spacing": "12px"}, "subwidgets": [ { - "widgettype": "urlwidget", - "id": "file_list_panel", - "options": { - "url": "{{entire_url('./file_list.dspy')}}?kb_id={{params_kw.kb_id}}&id=__root__" - } - } - ], - "binds": [ - { - "wid": "dir_tree", - "event": "node_selected", - "actiontype": "urlwidget", - "target": "file_list_panel", - "mode": "replace", - "options": { - "url": "{{entire_url('./file_list.dspy')}}?kb_id={{params_kw.kb_id}}" - } - } + "widgettype": "UiFile", "id": "file_drop_zone", + "options": {"accept": "", "multiple": true, "preview": false, + "otext": "📂 拖拽文件到此处上传，或点击选择文件", + "width": "100%", "cheight": 10, "css": "card", + "border": "2px dashed #3b82f6", "alignItems": "center", "justifyContent": "center", + "kb_id": "{{params_kw.kb_id}}"}, + "binds": [ + {"wid": "self", "event": "dragover", "actiontype": "script", + "script": "event.preventDefault();event.stopPropagation()"}, + {"wid": "self", "event": "drop", "actiontype": "script", + "script": "event.preventDefault();event.stopPropagation()"} + ] + }, + {"widgettype": "Text", "id": "upload_status", + "options": {"cfontsize": 12, "color": "#50b86c", "marginTop": "0", "text": "支持 TXT/PDF/JPG/PNG/MP3/WAV/MP4"}}, + {"widgettype": "VScrollPanel", "id": "file_list_panel", + "options": {"css": "filler", "padding": "0", "spacing": "0"}, + "subwidgets": [ + {"widgettype": "urlwidget", + "options": {"url": "{{entire_url('./file_list.dspy')}}?kb_id={{params_kw.kb_id}}", "css": "filler"}} + ]} ] } ] } + ], + "binds": [ + {"wid": "dir_tree", "event": "node_selected", "actiontype": "script", + "script": "window._rag_folder=event.params.id||''"}, + {"wid": "file_drop_zone", "event": "changed", "actiontype": "script", "target": "file_drop_zone", + "script": "var kb_id=this.opts.kb_id||'';var tree=bricks.getWidgetById('dir_tree',bricks.app.root);var folder=window._rag_folder||'';var files=this.value;if(!files)return;if(!Array.isArray(files))files=[files];var st=document.getElementById('upload_status');var done=0;var total=files.length;function refreshList(){var fp=bricks.getWidgetById('file_list_panel',bricks.app.root);fp.clear_widgets();var url=location.origin+'/rag/knowledge_bases_list/file_list.dspy?_webbricks_=1&kb_id='+kb_id+(folder?'&id='+folder:'');bricks.widgetBuild({widgettype:'urlwidget',options:{url:url,css:'filler'}},fp).then(function(w){if(w)fp.add_widget(w)})};for(var i=0;i=total)refreshList()};x.onerror=function(){done++;if(st)st.innerText='上传失败';if(done>=total)refreshList()};x.send(f)})(files[i])}"}, + {"wid": "dir_tree", "event": "node_selected", "actiontype": "urlwidget", "target": "file_list_panel", "mode": "replace", + "options": {"url": "{{entire_url('./file_list.dspy')}}?kb_id={{params_kw.kb_id}}"}} ] } diff --git a/wwwroot/knowledge_bases_list/upload_file.dspy b/wwwroot/knowledge_bases_list/upload_file.dspy new file mode 100644 index 0000000..a0184e7 --- /dev/null +++ b/wwwroot/knowledge_bases_list/upload_file.dspy @@ -0,0 +1,222 @@ +ns = params_kw.copy() +kb_id = ns.get('kb_id', '') +folder_id = ns.get('folder', '') +file_name = ns.get('file_name', 'upload.bin') +if not kb_id: + return json.dumps({"status": "error", "error": "kb_id required"}, ensure_ascii=False) +file_data = await request.read() +if not file_data: + return json.dumps({"status": "error", "error": "no file data"}, ensure_ascii=False) + +env = request._run_ns +userorgid = await env.get_userorgid() +doc_id = str(uuid()).replace('-', '')[:16] +ext = '.' + file_name.rsplit('.', 1)[1] if '.' in file_name else '.bin' +file_path = '/d/rag/ragserver/pkgs/rag/rag/files/' + doc_id + ext +with open(file_path, "wb") as f: + f.write(file_data) +file_size = len(file_data) + +ext_l = ext.lower() +text_exts = {'.txt', '.md', '.csv', '.json', '.xml', '.html', '.htm', '.py', '.js', '.css', '.yaml', '.yml', '.log', '.rst'} +doc_exts = {'.pdf', '.docx', '.pptx', '.xlsx', '.doc', '.ppt', '.xls'} +image_exts = {'.jpg', '.jpeg', '.png', '.bmp', '.gif', '.webp'} +audio_exts = {'.mp3', '.wav', '.flac', '.ogg', '.m4a', '.aac'} +video_exts = {'.mp4', '.avi', '.mov', '.mkv', '.webm'} + +text = '' +chunks_n = 0 +face_count = 0 +voice_speakers = 0 +meta_parts = {} + +# --- TEXT EXTRACTION --- +if ext_l in text_exts: + text = file_data.decode('utf-8', errors='replace') +elif ext_l == '.pdf': + import io; from PyPDF2 import PdfReader + try: + r = PdfReader(io.BytesIO(file_data)) + text = '\n'.join(p.extract_text() or '' for p in r.pages) + except: pass +elif ext_l == '.docx': + import io; from docx import Document + try: + doc = Document(io.BytesIO(file_data)) + text = '\n'.join(p.text for p in doc.paragraphs if p.text.strip()) + except: pass +elif ext_l == '.pptx': + import io; from pptx import Presentation + try: + prs = Presentation(io.BytesIO(file_data)) + parts = [] + for s in prs.slides: + for sh in s.shapes: + if hasattr(sh, 'text') and sh.text.strip(): + parts.append(sh.text) + text = '\n'.join(parts) + except: pass +elif ext_l in {'.xlsx', '.xls'}: + import io; from openpyxl import load_workbook + try: + wb = load_workbook(io.BytesIO(file_data), read_only=True, data_only=True) + rows = [] + for ws in wb.worksheets: + for row in ws.iter_rows(values_only=True): + r = ' | '.join(str(c) if c is not None else '' for c in row) + if r.strip(): rows.append(r) + text = '\n'.join(rows) + except: pass + +import aiohttp, base64 + +# --- IMAGE: face detection + CLIP visual embedding --- +if ext_l in image_exts: + img_b64 = base64.b64encode(file_data).decode() + async with aiohttp.ClientSession(timeout=aiohttp.ClientTimeout(total=15)) as s: + # Face detection + try: + r = await s.post('https://media.opencomputing.net:10443/face/api/detect', json={"images": [img_b64]}) + if r.status == 200: + fd = await r.json() + results = fd.get("results", []) + if results and isinstance(results[0], dict): + face_count = len(results[0].get("faces", results[0].get("detections", []))) + meta_parts['face'] = face_count + except: pass + +# --- AUDIO: voiceprint extraction --- +if ext_l in audio_exts: + async with aiohttp.ClientSession(timeout=aiohttp.ClientTimeout(total=60)) as s: + try: + form = aiohttp.FormData() + form.add_field('file', file_data, filename=file_name) + r = await s.post('https://media.opencomputing.net:10443/voiceprint/extract/submit', data=form) + if r.status == 200: + vd = await r.json() + if vd.get('status') == 'SUCCEEDED': + voice_speakers = vd.get('speakers', 1) + elif vd.get('embedding'): + voice_speakers = 1 + meta_parts['voiceprint'] = voice_speakers + except: pass + +# --- VIDEO: frame extraction + audio extraction (via ffmpeg) --- +if ext_l in video_exts: + import subprocess, tempfile + meta_parts['video'] = 'pending' + # Extract keyframe for face detection + try: + tmp_img = '/tmp/' + doc_id + '_frame.jpg' + subprocess.run(['ffmpeg', '-y', '-i', file_path, '-vframes', '1', '-q:v', '2', tmp_img], + capture_output=True, timeout=30) + if __import__('os').path.exists(tmp_img): + with open(tmp_img, 'rb') as fi: + frame_data = fi.read() + img_b64 = base64.b64encode(frame_data).decode() + async with aiohttp.ClientSession(timeout=aiohttp.ClientTimeout(total=15)) as s: + try: + r = await s.post('https://media.opencomputing.net:10443/face/api/detect', json={"images": [img_b64]}) + if r.status == 200: + fd = await r.json() + results = fd.get("results", []) + if results and isinstance(results[0], dict): + face_count = len(results[0].get("faces", results[0].get("detections", []))) + except: pass + __import__('os').remove(tmp_img) + except: pass + # Extract audio for voiceprint + try: + tmp_audio = '/tmp/' + doc_id + '_audio.wav' + subprocess.run(['ffmpeg', '-y', '-i', file_path, '-vn', '-acodec', 'pcm_s16le', '-ar', '16000', '-ac', '1', tmp_audio], + capture_output=True, timeout=60) + if __import__('os').path.exists(tmp_audio) and __import__('os').path.getsize(tmp_audio) > 44: + with open(tmp_audio, 'rb') as fa: + audio_data = fa.read() + async with aiohttp.ClientSession(timeout=aiohttp.ClientTimeout(total=60)) as s: + try: + form = aiohttp.FormData() + form.add_field('file', audio_data, filename='audio.wav') + r = await s.post('https://media.opencomputing.net:10443/voiceprint/extract/submit', data=form) + if r.status == 200: + vd = await r.json() + if vd.get('status') == 'SUCCEEDED' or vd.get('embedding'): + voice_speakers = vd.get('speakers', 1) + except: pass + __import__('os').remove(tmp_audio) + except: pass + meta_parts['face'] = face_count + meta_parts['voiceprint'] = voice_speakers + +# --- RAG INGEST (text + image CLIP embedding) --- +if text and len(text.strip()) > 10: + paragraphs = text.split('\n') + chunks = [] + cur = '' + for p in paragraphs: + p = p.strip() + if not p: + if cur: chunks.append(cur); cur = '' + continue + if len(cur) + len(p) < 500: + cur = (cur + '\n' + p).strip() + else: + if cur: chunks.append(cur) + cur = p + if cur: chunks.append(cur) + + if chunks: + async with aiohttp.ClientSession(timeout=aiohttp.ClientTimeout(total=15)) as s: + try: + r = await s.post('https://embedding.opencomputing.net:10443/api/embed', + json={"texts": chunks, "model": "CLIP-ViT-H-14"}) + emb_data = await r.json() if r.status == 200 else {} + embeddings = emb_data.get("text_embeddings", emb_data.get("embeddings", [])) + except: + embeddings = [] + + vector_ids = [] + if embeddings: + try: + vdb_data = {"collection": kb_id, "data": [ + {"id": doc_id + "_" + str(i), "vector": emb, "text": chunks[i]} + for i, emb in enumerate(embeddings)]} + await s.post('https://vectordb.opencomputing.net:10443/v1/upsert', json=vdb_data) + vector_ids = [doc_id + "_" + str(i) for i in range(len(embeddings))] + except: + pass + + async with get_sor_context(env, 'rag') as sor: + for i, chunk_text in enumerate(chunks): + vid = vector_ids[i] if i < len(vector_ids) else '' + await sor.sqlExe( + "INSERT INTO document_chunks (id, doc_id, kb_id, chunk_index, content, vector_id, created_at) " + "VALUES (" + "${id}$, ${doc_id}$, ${kb_id}$, ${idx}$, ${content}$, ${vid}$, NOW())", + {"id": doc_id + "_c" + str(i), "doc_id": doc_id, "kb_id": kb_id, + "idx": i, "content": chunk_text[:2000], "vid": vid}) + chunks_n = len(chunks) + +# --- SAVE TO DB --- +import json as _json +meta_json = _json.dumps(meta_parts, ensure_ascii=False) +status = 'done' +async with get_sor_context(env, 'rag') as sor: + await sor.sqlExe( + "INSERT INTO documents (id, kb_id, folder_id, file_name, file_type, file_size, file_path, mime_type, status, chunk_count, metadata, org_id, created_at, updated_at) " + "VALUES (" + "${id}$, ${kb_id}$, ${folder_id}$, ${file_name}$, 'other', ${file_size}$, ${file_path}$, 'application/octet-stream', ${status}$, ${chunks}$, ${meta}$, ${org_id}$, NOW(), NOW())", + {"id": doc_id, "kb_id": kb_id, "folder_id": folder_id, "file_name": file_name, + "file_size": file_size, "file_path": "/idfile/files/" + doc_id + ext, + "status": status, "chunks": chunks_n, "meta": meta_json, "org_id": userorgid}) + await sor.sqlExe( + "UPDATE knowledge_bases SET doc_count=doc_count+1, total_size=total_size+" + "${size}$ WHERE id=${kb_id}$", + {"size": file_size, "kb_id": kb_id}) + if chunks_n: + await sor.sqlExe( + "UPDATE knowledge_bases SET chunk_count=chunk_count+" + "${n}$ WHERE id=${kb_id}$", + {"n": chunks_n, "kb_id": kb_id}) + +result = {"status": "SUCCEEDED", "doc_id": doc_id, "file_name": file_name, + "file_size": file_size, "folder_id": folder_id, + "text_len": len(text), "chunks": chunks_n, + "faces": face_count, "speakers": voice_speakers} +return _json.dumps(result, ensure_ascii=False, default=str)